add site GET parameter

diff --git a/.gitignore b/.gitignore
index d041e45b2cbf0963e7f61e4b2ec62521f93436a1..bda5eebb8f002edd83d3c60f727679fbfebddd83 100644 (file)
--- a/.gitignore
+++ b/.gitignore
@@ -1 +1,2 @@
 /data/config.php
 /data/config.php

+README.html

diff --git a/README.rst b/README.rst
index ba8a681e57bae93488ab782876c8a2e8c31b6af2..f1e0abb06d0eba82aa11487b5c268932e37bf275 100644 (file)
--- a/README.rst
+++ b/README.rst
@@ -1,8 +1,18 @@
 Features
 ========
 Features
 ========

+- Crawler and indexer with the ability to run many in parallel
+- Shows and highlights text that contains search words
+- Boolean search queries:
+
+  - ``foo bar`` searches for ``foo AND bar``
+  - ``foo OR bar``
+  - ``title:foo`` searches for ``foo`` only in the page title
+- Facets for tag, domain, language and type

 - Site search
 
 - Site search
 

-  - Query: ``foo bar site:www.example.org/dir/``
+  - Query: ``foo bar site:example.org/dir/``
+  - or use the ``site`` GET parameter:
+    ``/?q=foo&site=example.org/dir``

 
 Dependencies
 ============
 
 Dependencies
 ============

diff --git a/data/templates/search.htm b/data/templates/search.htm
index 4f3a252ace8e4262bcc1503fdadf474c1768ae2e..d7037847867da7dc5475aea641b589506272d47a 100644 (file)
--- a/data/templates/search.htm
+++ b/data/templates/search.htm
@@ -14,6 +14,9 @@
      <form class="navbar-form pull-left">
       <input type="text" name="q" placeholder="Search"
              value="{{query}}" class="input-xxlarge"/>
      <form class="navbar-form pull-left">
       <input type="text" name="q" placeholder="Search"
              value="{{query}}" class="input-xxlarge"/>

+      {% if siteParam %}
+      <input type="hidden" name="site" value="{{site}}"/>
+      {% endif %}

       <button type="submit" class="btn">Find</button>
      </form>
     </div>
       <button type="submit" class="btn">Find</button>
      </form>
     </div>

diff --git a/data/templates/search/list.htm b/data/templates/search/list.htm
index 46bf4bae8655f0449191927f88b332c0919c46aa..069dcf3fce3714c07b9101406feba5644423026b 100644 (file)
--- a/data/templates/search/list.htm
+++ b/data/templates/search/list.htm
@@ -1,10 +1,16 @@
 {% if hitcount == 0 %}
  <p>
 {% if hitcount == 0 %}
  <p>

-  Sorry, no results for "<tt>{{query}}</tt>".
+  No results
+  for "<strong><tt>{{cleanQuery}}</tt></strong>"
+  {% if site %}
+  on <strong><tt>{{site}}</tt></strong>
+  &#160;&#160;&#160;|&#160;&#160;&#160;
+  <a href="{{urlNoSite}}">Show all results</a>
+  {% endif %}

  </p>
 {% else %}
  </p>
 {% else %}

- <p class="resultinfo">
-  {{hitcount}} results
+<p class="resultinfo">
+  {{hitcount}}

   {% if hitcount == 1 %}result{% else %}results{%endif%}
   for "<strong><tt>{{cleanQuery}}</tt></strong>"
   {% if site %}
   {% if hitcount == 1 %}result{% else %}results{%endif%}
   for "<strong><tt>{{cleanQuery}}</tt></strong>"
   {% if site %}

diff --git a/www/index.php b/www/index.php
index 846d589fef148bb0aacc7c30bcae9ae79c550a85..6f4ec8fd15f0286b535dd0ba9208060ceec307a5 100644 (file)
--- a/www/index.php
+++ b/www/index.php
@@ -17,9 +17,23 @@ if (isset($_GET['page'])) {
     $page = (int)$_GET['page'] - 1;
 }
 $perPage = 10;//$GLOBALS['phinde']['perPage'];
     $page = (int)$_GET['page'] - 1;
 }
 $perPage = 10;//$GLOBALS['phinde']['perPage'];

-
+$site = null;
+$siteParam = false;

 $baseLink = '?q=' . urlencode($query);
 
 $baseLink = '?q=' . urlencode($query);
 

+if (preg_match('#site:([^ ]*)#', $query, $matches)) {
+    $site = $matches[1];
+    $cleanQuery = trim(str_replace('site:' . $site, '', $query));
+    $site = Helper::noSchema($site);
+} else if (isset($_GET['site']) && trim(isset($_GET['site'])) != '') {
+    $site = trim($_GET['site']);
+    $siteParam = true;
+    $cleanQuery = $query;
+    $baseLink .= '&site=' . urlencode($site);
+} else {
+    $cleanQuery = $query;
+}
+

 $filters = array();
 if (isset($_GET['filter'])) {
     $allowedFilter = array('domain', 'language', 'tags', 'term');
 $filters = array();
 if (isset($_GET['filter'])) {
     $allowedFilter = array('domain', 'language', 'tags', 'term');


@@ -54,7 +68,6 @@ function buildLink($baseLink, $filters, $addFilterType, $addFilterValue)
     return $baseLink;
 }
 
     return $baseLink;
 }
 

-$site = null;

 if (preg_match('#site:([^ ]*)#', $query, $matches)) {
     $site = $matches[1];
     $cleanQuery = trim(str_replace('site:' . $site, '', $query));
 if (preg_match('#site:([^ ]*)#', $query, $matches)) {
     $site = $matches[1];
     $cleanQuery = trim(str_replace('site:' . $site, '', $query));


@@ -77,7 +90,8 @@ $pager = new Html_Pager(
 
 foreach ($res->hits->hits as &$hit) {
     $doc = $hit->_source;
 
 foreach ($res->hits->hits as &$hit) {
     $doc = $hit->_source;

-    if ($doc->title == '') {
+    if (!isset($doc->title) || $doc->title == '') {
+        $doc->title = '(no title)';

         $doc->htmlTitle = '(no title)';
     }
     if (isset($hit->highlight->title[0])) {
         $doc->htmlTitle = '(no title)';
     }
     if (isset($hit->highlight->title[0])) {


@@ -104,6 +118,12 @@ foreach ($res->aggregations as $key => &$aggregation) {
     }
 }
 
     }
 }
 

+if ($site !== null) {
+    $urlNoSite = buildLink('?q=' . urlencode($cleanQuery), $filters, null, null);
+} else {
+    $urlNoSite = null;
+}
+

 render(
     'search',
     array(
 render(
     'search',
     array(


@@ -112,6 +132,7 @@ render(
         'cleanQuery' => $cleanQuery,
         'urlNoSite' => $urlNoSite,
         'site' => $site,
         'cleanQuery' => $cleanQuery,
         'urlNoSite' => $urlNoSite,
         'site' => $site,

+        'siteParam' => $siteParam,

         'hitcount' => $res->hits->total,
         'hits' => $res->hits->hits,
         'aggregations' => $res->aggregations,
         'hitcount' => $res->hits->total,
         'hits' => $res->hits->hits,
         'aggregations' => $res->aggregations,